This research will examine the relationship between alcohol and homelessness within the context of a rigorously designed study. Specifically the project will examine: 1. The degree to which alcohol, independent of other presumed causes, can be linked to homelessness. 2. The factors that condition whether or not alcohol use results in homelessness. 3. The patterns of homelessness, measured over a period of six months, of individuals who vary in the severity of their drinking patterns. 4. The role played by social agencies and treatment programs in assisting homeless people who have drinking related problems, and the barriers to service usage. The analysis will be based on data recently collected in two major cities, Chicago and Minneapolis, using similar questionnaires. The Chicago data were obtained by interviewing 500 men and women, about half of whom were homeless. Base on probability sampling, it allows us to compare the homeless to a sensible control group. The Minneapolis data contain information on the homeless obtained at two points in time, enabling us to trace the dynamics of homelessness. Wave one data in Minneapolis were obtained in interviews with 450 homeless individuals; approximately 255 of these individuals were relocated and interviewed.